<h2></h2><p>A speech-to-video character animation workflow powered by HuMo-17B, turning a single still into a talking-head performance driven by uploaded audio. Audio is encoded with Whisper-large v3, merged with CLIP prompt conditioning and a reference image to build motion latents in WanHuMoImageToVideo. KSampler with SD3 sampling head refines lip motion and facial dynamics, and VideoHelperSuite muxes frames with the audio at 25 FPS.​</p><h2 id="key-features">Key features</h2><ul><li><p>HuMo-17B GGUF backbone with SD3 sampling head for stable face dynamics and expressive lip sync.​</p></li><li><p>Whisper-large v3 audio encoder for timing-aware mouth shapes and speech rhythm.​</p></li><li><p>All-in-one subgraph encapsulating text prompt, dimensions, model names, and sampler settings for clean UX.​</p></li></ul><h2 id="inputs">Inputs</h2><ul><li><p>Reference image loader, audio upload (MP3/WAV), positive/negative prompts for scene, style, and camera.​</p></li><li><p>Defaults: 448×576, ~177 frames, 25 FPS output MP4 with audio.​</p></li></ul><h2 id="recommended-settings">Recommended settings</h2><ul><li><p>Sampler: UniPC, steps 4, denoise 1.0; SD3 sampling head scale 5.​</p></li><li><p>LoRA: LightX2V at strength 1.1 for subtle motion enhancement without overbaking identity.​</p></li></ul><h2 id="where-each-model-goes">Where each model goes</h2><ul><li><p>Diffusion model (GGUF): HuMo-17b-Q8_0.gguf — loaded by LoaderGGUF as base model before SD3 sampling head. Place under models used by GGUF loader.​</p></li><li><p>VAE: pig_wan_vae_fp32-f16.gguf — loaded by VaeGGUF and used by VAEDecode and WanHuMoImageToVideo. Place under models/vae.​</p></li><li><p>CLIP text encoder: cow-umt5xxl-q8_0.gguf — loaded by ClipLoaderGGUF with vision “wan” and projection “default”. Place under models/clip.​</p></li><li><p>LoRA: wan_loras/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors — LoraLoaderModelOnly before ModelSamplingSD3, strength 1.1. Place under models/loras/wan_loras.​</p></li><li><p>Audio encoder: whisper_large_v3_fp16.safetensors — AudioEncoderLoader and AudioEncoderEncode for speech features. Place under models/audio or the audio encoder directory.​</p></li></ul><h2 id="node-highlights">Node highlights</h2><ul><li><p>WanHuMoImageToVideo: fuses CLIP conditioning, VAE, audio features, and reference image into motion latents.​</p></li><li><p>ModelSamplingSD3 + KSampler: sets SD3 sampling parameters, then denoises with prompt control for identity stability.​</p></li><li><p>VHS_VideoCombine: assembles frames and the original audio into an H.264 MP4 at 25 FPS with low CRF.​</p></li><li><p>Subgraph IO: exposes width, height, length, gguf_name, lora_name, clip_name, vae_name, audio_encoder_name, sampler/scheduler, and seed for quick reuse</p></li></ul><p></p>

Files with SHA256: 9379eebcc82dee7a933d1f53d6321db671ce555c49e3ded9c943c00da3fb2241